Volume 224, Issue 2 p. 987-993
Free Access

Fiber-optic refractometer for in vivo sugar concentration measurements of low-nectar-producing flowers

Giovanna Aronne

Corresponding Author

Giovanna Aronne

Department of Agricultural Sciences, University of Naples Federico II, via Università 100, 80055 Portici (NA), Italy

Author for correspondence:

Giovanna Aronne

Tel: +39 081 253 9443

Email: [email protected]

Search for more papers by this author
Pietro Malara

Pietro Malara

CNR-Istituto Nazionale di Ottica, via Campi Flegrei 34 (comprensorio A. Olivetti), 80078 Pozzuoli (NA), Italy

Search for more papers by this author
First published: 31 July 2019
Citations: 12


  • Sugar concentration in floral nectars is an assessment required in several diverse fields of application. The widely used analysis, consisting of nectar extraction with a microcapillary and sugar concentration measurement with a light refractometer, is not reliable when the nectar is secreted in small quantities, unextractable with a microcapillary. Ancillary methods adopted in such cases are destructive, rather complicated and often provide much less precise and accurate results.
  • The microscopic-size, low cost and biocompatibility of optical fibers were exploited to deliver light directly inside the flower with minimal invasiveness and measure instantaneously the refractometric properties of the nectar without extracting it. After comparing the new and old methods using two known nectariferous species, the new approach was validated on Primula palinuri, whose nectar is unextractable with microcapillaries.
  • The fiber-optic probe was able to measure the nectar refractive index in P. palinuri flowers making it possible to highlight a previously undetected significant trend of the sugar concentration throughout the long anthesis of the single flowers. Changes in nectar concentrations are similar in both longistylous and brevistylous flowers.
  • The fiber-optic refractometer is an advancement of light refractometer analysis. Further customization of the laboratory set-up into portable equipment will boost applications.


The enormous diversity of flower traits evolved within the angiosperms also includes a large variability of nectar production, both in terms of quality and quantity. Interest in flower nectar production is currently increasing and research is aimed at achieving very diverse goals: these include the evolution of plant–pollinator interactions (e.g. Nepi et al., 2008; Klumpers et al., 2019; Parachnowitsch et al., 2019), predictive effects of climate changes (e.g. Phillips et al., 2018; Takkis et al., 2018), conservation of endangered species (e.g. Montagna et al., 2018; Teixeira et al., 2018) and enhancement in food production (e.g. Prasifka et al., 2018; Soto et al., 2018).

Nectar is a water solution containing sugars, but also amino acids, proteins, inorganic ions and other very minor amounts of compounds such as lipids, phenolics and alkaloids (Nicolson & Thornburg, 2007). Sugars represent by far the most abundant nectar compounds; their concentration ranges between 10% and 75% (Willmer, 2011). Glucose, fructose and sucrose are the three main nectar sugars; their proportions vary according to species and are reported as adaptive traits to different pollinator types (Baker & Baker, 1983). A selective pressure on pollinator types also is enacted by nectar viscosity, because stickiness limits the ability to extract the nectar. As a rule-of-thumb, viscosity increases exponentially with total sugar concentration, independently of the proportions of different sugars. Studies on constancy of sugar composition have highlighted that these proportions change with flower age and vary among flowers of the same plant and of different plants due to several phenomena including nectar fermentation and interaction with environmental factors (e.g. Nicolson & van Wyk, 1998; Nepi, 2017). Despite the complex interactions among these variables, the total sugar content of nectar remains the most suitable quantitative indicator of flower attractiveness to pollinators, whereas more detailed analyses are performed only to address specific research questions (Nepi, 2014).

Total sugar concentration in nectar is generally assessed by extracting the liquid from the flower with a volumetric glass microcapillary tube, spilling the drop on the prism of a light refractometer and measuring its refractive index. The refractive index of nectar is used as a measure of sucrose equivalents and is based on the assertions that glucose and fructose have similar refractive indices (Corbet, 1978; Baker & Baker, 1982), which in turn are approximately half that of equimolar sucrose (Hainsworth & Wolf, 1972). The percentage of sucrose equivalents needs to be corrected for temperature and can be measured with different types of refractometers. Some of them cover the whole range of concentration, others, more accurate, measure only within shorter intervals (0–32% and 28–62% are the most widely used for nectar analysis). To ease the calculation of the quantity of sugar in the nectar, most light refractometers measure the percentage sucrose on the BRIX scale (1°Bx is 1 g of sucrose in 100 g of solution).

The amount of nectar produced in a flower varies according to plant taxonomy and to selective pressure exerted by pollinator types. For bee-pollinated flowers it is reported to vary from 0.1 to 10 μl (Opler, 1983). Flowers containing < 0.5 μl of nectar are considered low-nectar-producing. Some refractometers can be adapted for these working conditions (Dafni, 1992), but when the nectar content is further reduced to c. 0.1 μl, both the extraction and the refractometric measurement are less effective (McKenna & Thomson, 1988; Mallick, 2000; Manetas & Petropoulou, 2000; Corbet, 2003). The occurrence of such a limiting constraint is far from an exceptional phenomenon. Species with similar nectar crop can be characterized by few flowers with abundant nectar production or numerous flowers with limited amount of nectar. In the latter case, tens or hundreds of flowers are often bunched in inflorescences (e.g. capitula, spikes, umbels); although the quantity of nectar per single flower is insignificant, overall the inflorescence produces an amount that greatly attracts insect pollinators. Low-nectar-producing flowers are related not only to plant taxonomy, but also to habitat peculiarities. For instance, it is known that low-nectar-producing species are particularly abundant in Mediterranean communities (Petanidou & Smets, 1995), a feature that is usually linked to the soil drought and the high temperatures occurring in this geographical area during the main flowering period (from Spring to Summer).

In order to overcome technical constraints of sugar concentration analysis in low-nectar-producing flowers, several methods, ancillary to microcapillary extraction and subsequent light refractometer analysis, have been proposed (Kearns & Inouye, 1993; Dafni et al., 2005). In most cases, attention is focused on the extraction procedure. For instance, to enhance the performance of the glass microcapillary, Corbet (2003) suggests to preliminarily melt the midpart of the tube with a flame, to stretch it into a fine hair-like strand by pulling the two ends apart and to break the tube in two microcaps. In such cases, nectar extraction is better performed under a stereomicroscope to avoid piercing the floral tissue with the sharp tip and clogging the very thin lumen. Generally, with this method several flowers are necessary to extract enough droplets to be deposited on the refractometer prism before amassing sufficient volume to be analyzed.

An alternative suggestion for nectar extraction is the use of filter paper wicks. A small triangle of filter paper is held with forceps and the narrow corner placed on to the nectary to absorb the nectar. The paper wicks are then soaked in a vial with a known volume of distilled water and the original concentration of the nectar is calculated from that of the water solution (e.g. McKenna & Thomson, 1988). Both of these methods are unsuitable when nectar is too viscous. In such cases, it is possible to water-solve the nectar in the flower with different procedures of washing or rinsing (Morrant et al., 2009) and use the solution to work out sugar concentration. The above-mentioned methods are less practical than that attainable by using microcapillary extraction and directly dropping the nectar onto the light refractometer prism.

Within this framework, the aim of the current research activities was to find a method based on the refractive index principle which was able to measure total sugar concentration in low-nectar-producing flowers while avoiding problems related to the extraction.

The refractive index sensing capabilities of optical fibers have been known for long time (Meyer & Easley, 1987). In the last two decades, the evolution of lithographic fabrication techniques led to the development of sophisticated fiber-optic refractometers based on plasmonic or photonic nanostructures (Liang et al., 2005). Although these sensors have demonstrated ultra-high sensitivity and accuracy, they are extremely delicate and rarely suited for use beyond the laboratory (Malara et al., 2018). In the present work, a very simple and robust fiber-optic refractometer is applied to the analysis of flower's nectar. The microscopic size, low cost and biocompatibility of optical fibers make them ideal contact-probes able to measure the refractive index of the nectar directly inside the flower. The fiber-optic probe demonstrated herein is particularly suited for the analysis of low-producing flowers, as it overcomes the issues connected to the extraction and measurement of small nectar volumes. The refractive index measurement is based on the Fresnel reflection of guided light from the cleaved end facet of the optical fiber, which is placed in contact with the nectar. The procedure is therefore instantaneous, minimally invasive and, because of the extremely reduced dimensions of the reflecting surface, can be carried out successfully even with few-micron size nectar droplets.

In order to test the reliability of the above-described method, sugar concentration of the nectar of two known nectariferous species was compared as measured with the fiber-optic refractometer and the light refractometer. To validate the potential of the new approach, a species from the Mediterranean flora was used as model example. Within the current scenario of increasing temperature and aridity conditions in the Mediterranean area, conservationists are paying great attention to changes in quantity and quality of nectar production, and relative consequences on species competition for insect attraction (Petanidou, 2007). The Mediterranean Region is recognized as one of the world's Biodiversity Hotspots and also as one of the most threatened, mainly by human activities that are leading an ever-increasing number of plant species towards a high risk of extinction. One such plant is Primula palinuri, a species reported as Endangered with in the IUCN (International Union for Conservation of Nature) Red List (Gangale et al., 2011) and relying on pollinators to develop seeds. Its flowers produce a small amount of viscous nectar that cannot be sampled using microcapillaries. Alternative nectar analysis methods, including filter-paper wicks and flower rinsing (Kearns & Inouye, 1993; Dafni et al., 2005), also proved to be unsuccessful for P. palinuri due either to the small sample volume and/or its high viscosity, making it impossible to evaluate nectar within the floral biology context of this species (Aronne et al., 2014).

Materials and Methods

Fiber-optic refractometer system

The setup for the in vivo nectar refractometric measurements is illustrated in Fig. 1. Radiation from a spontaneous emission near-infrared source (IPG ASE 100-C) is directed towards the cleaved end-facet of a standard telecommunications optical fiber (Corning® SMF-28®). To avoid damage to the end-facet, the radiation power injected in the fiber is kept below 5 mW. A fiber-optic circulator then directs the facet back-reflection towards a photodetector (Thorlabs PDA-255). The measurement is based on the Fresnel reflection principle, which states that the reflectivity R of a flat interface between two media with refractive indexes n1 and n2 is urn:x-wiley:0028646X:media:nph16084:nph16084-math-0001. In the present setup, the cleaved facet of the optical fiber represents an interface that separates the fiber glass material and the surrounding environment. When the fiber end is in contact with the nectar, the back-reflected intensity is
urn:x-wiley:0028646X:media:nph16084:nph16084-math-0002(Eqn 1)
(I0, incident intensity; nf and nnect, refractive indexes of the fiber glass material and the nectar, respectively; K, a factor that accounts for the back-coupling efficiency into the fiber-guided mode, and depends on the flatness of the fiber end-facet). In order to retrieve the nectar concentration nnect, IR was not measured directly, but rather its relative variation with respect to urn:x-wiley:0028646X:media:nph16084:nph16084-math-0003 – thus, the back-reflection intensity when the fiber is immersed in a bidistilled water reference:
urn:x-wiley:0028646X:media:nph16084:nph16084-math-0004(Eqn 2)
where nfiber and urn:x-wiley:0028646X:media:nph16084:nph16084-math-0005 are 1.44 and 1.33, respectively. Extracting the nnect from urn:x-wiley:0028646X:media:nph16084:nph16084-math-0006, rather than from the more direct IR signal allows mitigation of two important sources of measurement uncertainty. Indeed, as can be seen from Eqn Eqn 2:
  • urn:x-wiley:0028646X:media:nph16084:nph16084-math-0007 does not depend on I0, so its measurement is unaffected by the long-term power drifts of the radiation source.
  • urn:x-wiley:0028646X:media:nph16084:nph16084-math-0008, does not depend on K. This is essential for the consistency of recorded data, because the fiber facet must be cleaved after each measurement to remove the sticky nectar residues, and each cleaving results in a different value of K.
The fiber-optic refractometer used herein was calibrated by immersing the probe in solutions of bidistilled water and chemical-grade sucrose of known concentration, and recording for each solution the urn:x-wiley:0028646X:media:nph16084:nph16084-math-0009 signal. Fitting these data with a polynomial function returns a calibration relationship that allows direct conversion of the refractometer urn:x-wiley:0028646X:media:nph16084:nph16084-math-0010 signal into equivalent sucrose concentration units (BRIX). The calibration data, plotted along with their polynomial fit in Fig. 1, show that the instrument can correctly measure sucrose concentrations even close to the solubility limit at the laboratory temperature (23°C, c. 67.5 BRIX).
Details are in the caption following the image
Experimental setup and criterion for the division of flowers into the aging categories of young (A), mature (B) and senescent (C) (a); signal calibration with known-concentration sucrose solutions: experimental points and 2° order polynomial fit (b). Raw data available in Supporting Information Table S1. One degree BRIX is 1 g of sucrose in 100 g of solution.

Biological validation

In order to validate the fiber-optic refractometer as a new method to measure the sugar concentration of nectar, a set of comparative measurements against a usual light refractometer was carried out using flowers of known nectariferous species. Then, the potential of the new method was verified by applying it to a species with flowers that produce small nectar volumes unextractable with a microcapillary for conventional analysis with a hand-held light refractometer.

For the first objective, two species were selected coflowering at Portici (40°48′42″N–14°20′38″E), in the Botanical Garden of the Department of Agricultural Science, University of Naples Federico II: Aloë arborescens Mill. and Salvia officinalis L. Aloë species are widely distributed in dry habitats (such as the South African Fynbos that are characterized by a Mediterranean climate) and are well-known to have flowers with copius amounts of nectar (Nicolson & Nepi, 2005; Symes & Nicolson, 2008). Salvia species produce nectariferous flowers in the shrublands of the Mediterranean region (Dafni et al., 1988; Petanidou & Vokou, 1993; Petanidou et al., 2000; Mačukanović-Jocić et al., 2011).

For both species, a sample population of ten flowers at their first day of anthesis was considered. For each flower, the nectar was first analyzed by inserting the optical fiber into the bottom of the corolla tube. Contact with the nectar was signaled by an abrupt change of the back-reflection intensity. After the fiber measurement, the flower nectar was extracted with a microcapillary and dropped on the prism of a hand-held light refractometer (Atago N1, 0.0–32.0% Brix ATC) for a comparative measurement. The consistency of the two methods was ascertained by the superposition of respective mean concentration values and by the correlation between the fiber-optic and the usual light refractometer datasets.

A second experiment used as model species Primula palinuri, the only Mediterranean and maritime species of the whole Primula genus. It is a rare and endangered species surviving only in a narrow strip of c. 80 km along the Tyrrhenian coast of southern Italy. As in other species of the genus, P. palinuri is characterized by distyly, and therefore plants with brevistylous flowers and plants with longistylous flowers occur together in populations (Aronne et al., 2013). Flowers develop from the center of umbel inflorescences and, as they grow, their orientation changes from pendant to vertical. As shown in Fig. 1, three subsequent orientation stages are reported: (1) pendulous flower (180–120°); (2) horizontal flower (120–60°) and (3) vertical flower (60–0°). Different flower orientations correspond to different biological stages (young, mature or senescent, respectively) and were observed to exert different attractions to insect pollinators (Aronne et al., 2014). At any stage, P. palinuri flowers produce a very small amount of nectar, observable only under a microscope. The nectar droplet is secreted on the top of the ovary at the base of the stylus (Fig. 2). The present experiment used a collection of several individual plants for ex situ conservation and study purposes present in the Botanical Garden at Portici. Considering the rarity of the species and the limited number of individuals growing ex situ, for the present tests three different plants from each of the two morphs were used, grown in pots. At the time of full bloom, plants were transported from the Botanical Garden to the CNR laboratory and kept there just long enough for the measurements.

Details are in the caption following the image
Microscope images of Primula palinuri flowers with a fiber-optic probe and/or a microcapillary tube. (a) A brevistylous flower whose calix and corolla have been cut longitudinally and removed to compare size of the floral parts, a fiber-optic probe (left) and a microcapillary tube (right); (b) close-up of the pistil and of a fiber-optic probe (left) and a microcapillary tube (right); (c) nectar droplet on the nectary surface at the base of the style; (d) fiber-optic probing a nectar droplet. Bars, 1 mm.

Flowers were categorized as longistylous or brevistylous and classified in three subcategories: A (young), B (mature) and C (senescent) according to the biological stage inferred from its orientation (Fig. 1). Nectar concentration was measured in a total of 68 flowers. The collected datasets were analyzed statistically by using the OriginLab one-way ANOVA routine and Bonferroni post-hoc tests, to verify: (1) the equivalence of sugar concentration in longistylous and brevistylous flowers, and (2) the occurrence of a possible trend of sugar concentration during the flower life span in both morphs.


The A. arborescens and S. officinalis data show that nectar concentration measurements with the fiber-optic refractometer are statistically equivalent to those obtained with the usual light refractometer. Results are illustrated in Fig. 3. The mean concentration values obtained with the two methods (respectively) were fully consistent: 16.2 ± 0.5 vs 16.1 ± 0.5 BRIX for A. arborescens and 20.8 ± 0.3 vs 20.8 ± 0.2 for S. officinalis. The correlations between the fiber-optic and the refractometer data also were significant for both species (Pearson population and sample correlation coefficients: P = 0.86, r = 0.0012 for the A. arborescens data; P = 0.89, r = 0.0011 for the S. officinalis measurements).

Details are in the caption following the image
Equivalent sucrose concentration of nectar of Aloë arborescens and Salvia officinalis measured with the fiber-optic refractometer system (blue) and a hand-held light refractometer (orange). Data refer to means ± SE of the means. Raw data available in Supporting Information Table S2. One degree BRIX is 1 g of sucrose in 100 g of solution.

Results regarding P. palinuri showed that the fiber-optic refractometric system succeeded in measuring sugar concentration of the low-nectar-producing flowers. Although the 68-flower population available was not equally distributed between the two morphs and the three categories of anthesis stages, values of nectar concentrations (in BRIX units) could be compared between longistylous and brevistylous flowers in each of the three phenological stages (Fig. 4). The mean concentration values were consistent between longistylous and brevistylous flowers and remained in the 53–60 BRIX range during the flower life cycle. The data also highlight that within each morph, sugar concentration varied throughout anthesis, showing a distinct sugar peak in stage B (Fig. 4).

Details are in the caption following the image
Equivalent sucrose concentrations of nectar in the longistylous (green) and brevistylous (red) flowers of Primula palinuri in the three aging categories: A (young), B (mature) and C (senescent). Data refer to means ± SE of the means. Raw data available in Supporting Information Table S3. One degree BRIX is 1 g of sucrose in 100 g of solution.

Pooling together data of the two morphs, the mean sucrose concentration of the Ppalinuri nectar was assessed as 53.6 ± 0.6 (mean ± SE) Brix for A flowers, 59.0 ± 0.4 (mean ± SE) for B flowers and 56 ± 1 (mean ± SE) for C flowers (Table 1). A post hoc analysis of variance confirmed that the sucrose peak in B flowers was statistically significant. In particular, although pairwise mean-value comparison (Bonferroni test with confidence level 0.05) revealed that the difference between the concentrations of A and C flowers may have had a statistical origin (AC significance = 0), the largest value of the mature B flowers is to be considered a real feature (AB and BC significance = 1).

Table 1. Statistical comparison of nectar sucrose concentration of flowers of Primula palinuri throughout anthesis.
Flower stage Mean SE SD n
Global statistics
A 53.613 0.602 3.127 27
B 59.032 0.435 2.132 24
C 55.763 1.033 4.261 17
Mean diff SE P-value Significance
Bonferroni pairwise comparison (confidence level = 0.05)
AB 5.419 0.887 1.869E−7 1
AC 2.150 0.979 0.095 0
BC −3.269 1.002 0.005 1
  • Flower aging categories: A, young; B, mature; C, senescent. Data refer to means, SE, SD and number of samples (n). Bonferroni's pairwise comparison of the means returned a P-value much smaller than the confidence level (P = 0.05) for the AB and BC comparisons.


Quantity of nectar and sucrose equivalent concentration are among the most common traits used to characterize flower nectar production (Baker & Baker, 1983; Nicolson & Thornburg, 2007). The present data proved that sugar concentration measured with the fiber-optic instrument developed herein are consistent with usual light refractometric measurements. However, with the conventional technique, the assessment of sugar concentration is performed after extraction of nectar and measurement of its quantity. Therefore, where it is impossible to extract the nectar, it is impossible to analyze sugar concentration. The proposal herein to use the fiber-optic refractometer instead of the usual light refractometer allows measurement of sugar equivalent concentration without the need to extract nectar. The new approach therefore is especially valuable in those cases in which nectar production is so low that both quantity and quality are impossible to evaluate because nectar is unextractable.

One such a species is P. palinuri. Although long-term survival of this species relies on generation turnover and reproductive success depends on intermorph pollen transfer by insect pollinators, comparative analyses in nectar production had to be neglected in previous studies (Aronne et al., 2014) because previous attempts failed to measure sugar concentration of floral nectar.

Due to its small volume and high viscosity, the nectar of P. palinuri was difficult to extract with standard methods, but easily measurable with the fiber-optic refractometer. The diameter of the fiber optic is 125 μm, but the intensity of the guided light is concentrated in the fiber core. The area that needs to be in contact with the nectar has a diameter of only 10–15 μm, which is perfectly compatible with the size of the nectar droplets.

In P. palinuri, the performance of the fiber-optic refractometer exceeded expectation. Results allowed evaluation not only of average values of sugar concentration of the nectar within the species, but also of fine differences in nectar concentration between flower morphs and throughout their anthesis. From the species biology point of view, such findings confirmed the hypothesis that nectars of brevistylous and longistylous flowers have similar nectar concentration, therefore exerting the same attractiveness to insect pollinators (Aronne et al., 2014). Obtaining the same reward, insects visit the two flower morphs interchangeably favoring random distribution of pollen between the flowers. Data on nectar variability during anthesis, however, were even more useful for defining the flower biology of this rare species. Sugar concentration increased from flower opening to mature stage and subsequently decreased. These data are in line with and corroborate other results showing that insects visit mature flowers much more frequently than those in the other stages (Aronne et al., 2014). Moreover, the reduction of sugar concentration measured in senescent flowers fits with previous observations that their vertical position facilitates ingress of rainwater, in turn causing nectar dilution (Aronne et al., 2014).

Previous application of microcapillary extraction/light refractometer analysis and other ancillary methods to P. palinuri was further constrained by the limited number of available flowers due to the rarity of the species. High accuracy and precision, intrinsic to the fiber-optic refractometer, allowed reliable data to be obtained even using a limited number of flowers. Moreover, using the fiber-optic refractometer, no flower was destroyed because measurements were done in vivo. This can be considered an additional benefit for a species with conservation constraints.

Although the fiber-optic refractometer was tested with just one particularly rare and endangered Mediterranean species, in principle the method can be applied to any flower, including those such as Castanea sativa Miller with anemophilous characteristics (Giovanetti & Aronne, 2011). The possibility to analyze nectar instantaneously and in vivo could allow researchers to highlight additional patterns in nectar concentration without the constraints of microcapillary extraction. Further applications of this method may be envisaged to tackle issues not limited to species biology or biodiversity conservation, such as honey production and crop pollination services in the agricultural sector.

The assessment of nectar sugar concentration in continuum as well as in vivo has not been possible to date, even on abundantly producing flowers. The present method could address the needs of other research questions, including measurement of possible intrafloral variation of nectar sugar concentration due to floral aging, effective pollination events or environmental growing conditions (e.g. Canto et al., 2011), that have not been widely addressed so far because of the analytical limitations.

According to Amato & Petit (2017), the development of inexpensive and efficient methods is imperative to maximize accuracy of sugar measurements. The simplicity of operation, the reduced cost and the intrinsic ruggedness of the fiber-optic refractometer method make it a likely candidate to supercede the light refractometer method. At present this new approach to nectar analysis can be performed in any laboratory equipped with a fiber-optic refractometer. However, the impending customization of the laboratory setup in a portable instrument is highly likely to boost the application of this method for in situ measurements of nectar sugar concentration.

Author contributions

GA conceived the idea and planned the experiments; PM implemented the system and analyzed the data; and GA and PM equally contributed in performing the tests, interpreting the results and writing the paper.